Selecting the next WTO Director-General: What the trade community thinks
Matteo Fiorini, Bernard Hoekman, Petros Mavroidis, Douglas Nelson, Robert Wolfe 08 July 2020
On 14 May 2020, Roberto Azevedo, the Director-General of the WTO, announced he would step down from his position at the end of August. Governments are now in the process of selecting candidates for the position.1 In contrast to other international organisations, the leadership appointment process is open and relatively transparent: WTO members take the ultimate decision, based on consensus. As part of a broader research project on global trade governance at the European University Institute, we implemented a survey to solicit views on what WTO members and the international trade community consider the most important attributes of candidates for the position.
The survey ran between 5 June and 21 June 2020 on an online survey platform. The questionnaire was sent by email to all WTO delegations and distributed through email and social media channels to trade practitioners using a contact list developed by the Global Economics programme of the EUI’s Robert Schuman Centre for Advanced Studies. Recipients were requested to forward the survey to others in their networks working on trade and WTO matters. The sampling frame is not designed to ensure that respondents will be perfectly representative of the WTO stakeholder community, and no claim to this effect is made.
Responses to all questions used a five-point scale: very low (strongly disagree), low (disagree), neutral, high (agree) and very high (strongly agree). Respondents were asked to indicate their professional affiliation (more than one allowed), whether they are based in Geneva, their gender and the nationality of their company/organisation/institution, where relevant (no nationality information was requested from staff of the EU and international organisations).2 Any question could be left unanswered. The survey instrument and a more extensive presentation of results can be found in Fiorini et al. (2020).
A total of 1,092 people opened the survey, and 75% (819) responded to at least one question. Around 800 responses were received for most of the questions (the lowest number of respondents for a question is 733; the highest is 807). Government officials (including the EU) represent the second largest category of respondents (24.6%) after academia (25.1%). The private sector (companies and business associations) account for 19% of all respondents, followed by staff of international organizations (18%) and NGOs, labour unions, think tanks (11%). Of the government and EU respondents, 31% (66) are based in Geneva.
Preferred characteristics of the next DG
The responses ranking desirable professional characteristics of the next DG reveal three clear messages.
Message #1: Strong agreement on desirable professional attributes
Respondents gave the greatest weight to the following four attributes (ranked in order of the share of ‘high’ and ‘very high’ responses): (1) management experience; (2) political experience; (3) economics training; and (4) experience as a WTO negotiator. All four attracted more than 75% high/very high scores (Figure 1, panel a).
The nationality of the DG, the notion that he/she should come from a region that has not yet provided a DG, and the gender of the DG are less highly ranked. Fewer than 40% of respondents gave the latter two attributes a high or very high score, and more than 20% scored them as low or very low. The lowest-ranked attribute across the sample is whether the DG should be from a developed country: 90% of respondents rate this as very low, low or neutral. This is of interest given that some WTO members argue there is an informal understanding that the DG should rotate between a national from a developed country following one from a developing country. The incumbent is from an emerging economy.
Only 25% of the sample accord gender a rating of high or very high. Female respondents give this attribute a higher weighting than men, but do not differ in their overall ranking of the various characteristics.
Figure 1 Strong agreement on preferred attributes
(a) Attributes sorted by sum of high & very high shares
(b) Spearman rho correlation between group rankings
Note: The question underlying the correlations in panel (b) is: What priority do you assign to the following characteristics for the WTO DG? Number of respondents between parentheses. See Fiorini et al. (2020) for survey questions.
A ‘heatmap’ of group-specific Spearman rho correlations (Figure 1, panel b) reveals substantial overlap across groups of respondents regarding the ranking of the ten characteristics.3 Only two respondent groups differ significantly in their ranking of attributes: Africa and the Middle East.4 African respondents accord the greatest priority (highest ranking) for the DG being a national of developing country, as well as prioritising the notion of regional diversity. Almost 80% of African respondents rank order these two elements as high or very high.
Message #2: Competence and regional diversity over time
Overall, 70% of all respondents either agree (30%) or strongly agree (40%) that the selection of the next DG should be solely based on expertise and competence (Figure 2, panel a). At the same time, slightly more than 60% agree (43%) or strongly agree (20%) that the selection should ensure regional variation over time. Less priority is given to alternating (rotating) between a developed and a developing country national. Respondents also accord less support to the notion that the DG should not be from one of the three largest trading powers (China, the EU or the US). This last question generated a high share of very low or low responses (40%), indicating that while there is little support for the DG having to come from a developed country, there is relatively strong opposition to the notion that a DG should not come from one of the large trade powers. In conjunction with the discussion above, these results suggest a substantial preference for regional diversity, but less support for prejudging or restricting the origin of the DG.
Although there is broad agreement that competence is a high priority, there are also clear differences in rankings across some respondent groups. Geneva delegations place the greatest priority on competence as the dominant selection criterion, with more than 80% ranking this as high or very high. Conversely, African respondents do not rank competence first; the highest priority is accorded to regional diversity (more than 95% ranking this as high or very high) (Figure 2, panel b). Other groups of respondents that favour regional variation over time more strongly include non-Geneva-based government officials and respondents from NGOs and labour organizations.
Figure 2 The selection of the DG should…
(a) whole sample
(b) respondents from Africa
Note: Bars sorted by the joint share of agree and strongly agree responses.
Message #3: Connections to international organisations, major capitals and business matters most
Overall, there is strong agreement on assigning the highest priority to the new DG having personal connections and a professional network spanning: (i) international organisations; (ii) the capitals of the largest trade powers (Beijing; Berlin, Brussels (EU); Tokyo; and Washington DC); and (iii) international business. The share of high and very high priority responses for these three categories is above 60% for the total sample (Figure 3).5
Figure 3 How important are personal connections and recent professional experience with…
(a) Attributes sorted by sum of high & very high shares
(b) Spearman rho correlations between group rankings
Divergent rankings across groups
Estimation of ordered probit models reported in Fiorini et al. (2020) reveals some statistically significant divergences in rankings of desirable characteristics across groups.
- Capital-based government officials. Officials in capitals rate political experience and gender less highly than the total sample, while according a higher priority to WTO experience and coming from a developing country/ensuring regional diversity over time. This group of respondents also ranks links with international business, labour organisations, academia and think tanks more highly than the total sample.
- Geneva-based officials. Statistically significant differences in the (priorities) rankings of Geneva delegations are observed for the following attributes. Relative to the whole sample, Geneva delegations place higher priority on political experience, public profile, and connections in major capitals. Conversely, they rank developing country nationality, and coming from a region that has not yet supplied a DG, less highly than other respondents.
- Private sector and NGOs. Business is closely aligned with the whole sample but places a higher weight on management experience and, not surprisingly, having connections to international business. Relative to the overall sample, NGOs put greater priority on political experience, on the candidate originating in a developing country, on connections with academia, labour organisations, think tanks, and/or regional organisations, and on gender.
The responses to the survey are relatively clear. Competence, political experience, a network that spans major capitals, international business and international organisations, knowledge of the WTO negotiating process, and a background in economics are the characteristics that are ranked most highly. Moreover, Geneva-based officials often accord greater priority to these characteristics than the total sample does.
The responses to the survey suggest that the DG should not be a technocrat/WTO delegate. There is a recurring debate among trade officials and practitioners that has something of a cyclical dimension: should the WTO be led by an ex-minister/senior politician or by a bureaucrat (a trade official)? The most recent DG was a bureaucrat; some of his predecessors were former ministers or prime ministers (e.g. Supachai Panitchpakdi and Mike Moore). In 2013, most candidates had the political experience and type of network that many survey respondents rank highly (Hoekman and Mavroidis 2013). They were not selected. The survey responses suggest the pendulum may have swung back to a preference for the latter type of profile.
The converse of the high rank order given to the above-mentioned characteristics is that there is less support for the notion that the DG should come – or not come – from a developing/developed country; a region that has not yet supplied a DG; and that there should explicit rotation between DGs from the North and the South. There is however no consensus on this – African respondents rank these attributes much higher than the sample average. However, many groups also assign a high priority to regional diversity over time. This suggests that conditional on competence and the requisite political/professional networks, there is significant support for considering regional diversity as a factor – but not as a rule.
The priority accorded to the DG having connections with international organisations is striking. While it is impossible to make inferences as to why this is rated so highly, one hypothesis is that it may reflect a recognition that the trade agenda increasingly spans regulatory policies and an associated need for the WTO to bolster cooperation with agencies that have mandates and requisite expertise on such issues. As noted by Lamy (2013) when stepping down from the DG job, the WTO today has much more interactions with other international organizations, reflecting a recognition of the need for greater cooperation for policy coherence.
Beattie, A (2020), “Battle to head WTO offers chance to defend global trading order,” Financial Times, 15 June.
Fiorini, M, B Hoekman, P C Mavroidis, D Nelson and R Wolfe (2020), “Stakeholder Preferences and Priorities for the Next WTO Director General,” EUI RSCAS Working Paper 2020/43.
Hoekman, B and P C Mavroidis (eds) (2013), Race for the WTO Director-General job: Seven candidates speak, CEPR Press.
Lamy, P (2013), “Looking back, moving forward”, VoxEU.org, 29 July.
Reinsch, W (2020), “WTO: The Leadership Race Is On”, CSIS Commentary, 26 May.
1 See Reinsch (2020) and Beattie (2020) for discussion of the process and candidates that have been put forward.
2 Survey responses are anonymous; the software used makes it impossible for the researchers to identify respondents.
3 Spearman’s rho is a nonparametric method for measuring the correlation between two variables that are rank ordered. It ranges between -1 and 1. The higher the number the more similar is the rank ordering.
4 Respondents were mapped into country groupings using the WTO definition of geographic regions.
5 The high ranking accorded to links with international organizations is not influenced by a large share of respondents affiliated with such organizations. Both government officials and Geneva delegations rank links to international organizations as the top priority, while business respondents rank this as the third priority.